GPT Image 1 OpenAI vs Seedream 4.0 ByteDance

GPT Image 1

• + Excellent adherence to the lighting instruction with clear soft light from the left side.
• + Superior texture rendering on the book cover and the wooden table surface.
• + Clean, sharp glass edges and a more realistic plant appearance.

• − The blue sphere appears slightly matte and opaque, lacking the translucence one might expect from 'blue' in a glass setting.
• − The glass cube looks more like a frame than a solid cube.

Seedream 4.0

• + Beautiful light refraction and reflections on the glass and table surface.
• + The blue sphere has a realistic glass-like quality that complements the cube.
• + Accurately places the plant behind the cube as requested.

• − The book and plant lack the high-resolution detail found in the other image.
• − The structure of the glass cube's bottom is slightly messy and lacks clear geometry.

GPT Image 1

• + Excellent preservation of the specific Rolls Royce Phantom Drophead Coupe model.
• + High fidelity to the man's facial features and distinctive hairstyle.
• + Great sense of motion with realistic wheel blur and lighting.

• − The man appears slightly too large for the car interior.
• − Minor clipping where the hair meets the top of the windshield frame.

Seedream 4.0

• + Successfully captures the requested California coastline scenery and road layout.
• + Preserves the plaid pattern of the man's coat from the source image.
• + Good integration of the driver into the cockpit surroundings.

• − The man's face has changed significantly from the source image.
• − The car's proportions are slightly warped compared to the original source photo.
• − The road markings (double yellow lines) are crooked and inconsistent.

GPT Image 1

• + Excellent natural skin texture and facial details.
• + Mood lighting and shallow depth of field feel very professional and cinematic.
• + Good rendering of wet surfaces and raindrops on the jacket.

• − Missed the request for motion blur from passing cars (cars are static/out of focus).
• − The bicycle mechanics are a bit abstract and mushy upon closer inspection.

Seedream 4.0

• + Successfully captured the requested motion blur from passing cars.
• + Includes realistic elements like tools on the ground and a bicycle basket.
• + Accurate reflection on the wet pavement.

• − Image lacks the fine detail and sharpness; the man's face and hands are blurry.
• − Colors appear a bit washed out and less cinematic than Model A.
• − Anatomical issues with the fingers and how the man interacts with the bike.

GPT Image 1

• + Exceptional engraved plate armor detail and texture
• + Highly realistic skin rendering with natural-looking dirt and age
• + Atmospheric and moody lighting that highlights the metalwork

• − The beads in the hair are barely visible or too similar to hair color
• − The braided hair is a bit stiff/merged in some areas

Seedream 4.0

• + Strong prompt adherence regarding beads in the hair
• + Great inclusion of leather straps and cloth underlayer as requested
• + Genuinely lifelike and intense eyes

• − The torch in the background is a bit distracting and leads to blown-out highlights
• − Skin textures are slightly smoother and less 'battle-worn' than Model A

GPT Image 1

• + Clean grid layout that functions as an actual menu template
• + Consistent high-quality photography with white-background plates
• + Excellent text hierarchy and placeholder price alignment

• − Nonsense filler text used for descriptions
• − A 'Main' item is incorrectly listed under the 'Pizza' header

Seedream 4.0

• + Bold sans-serif typography as requested
• + Vibrant colors in the food photography

• − Lacks actual menu items and price list
• − Disorganized grid with overlapping or cut-off elements
• − Does not follow a professional menu layout

GPT Image 1

• + Excellent typography with clean, legible text and glowing effects.
• + Superior photorealistic detail in the food textures, especially the lettuce and patty.
• + Perfect adherence to the specified components and pricing layout.

• − The composition is a bit static and vertically symmetrical, reducing the sense of motion.
• − Did not include the '6' in the price starburst, rendering it as '€.99'.

Seedream 4.0

• + Strong sense of dynamic motion with swirling trails and a diagonal composition.
• + Creative interpretation of the fiery background with actual flames and embers.
• + Good integration of the text into the 3D space of the scene.

• − Failed the pricing prompt by rendering '€5.99' instead of '€6.99'.
• − The burger structure is conceptually messy with components duplicated or fused incorrectly.
• − Text rendering is slightly less crisp than the competitor's.

GPT Image 1

• + Excellent text legibility and spelling accuracy for all menu items.
• + Consistent chalk texture across all characters.
• + Clean composition that focuses entirely on the chalkboard.

• − The handwriting looks slightly digital/uniform, lacking the requested 'elegant cursive' for the title.
• − The framing is a tight crop that misses the 'cozy café' atmosphere mentioned in the prompt.

Seedream 4.0

• + Highly realistic 'handwritten' aesthetic with natural variations and authentic chalk smudges.
• + Successfully captures the 'cozy café' environment in the background.
• + Correctly interpreted the request for elegant cursive handwriting for the title.

• − Minor text artifacts like the hyphen before 'APRIL'.
• − Slightly less legible than Model A due to the authentic chalk dust and smearing effect.

GPT Image 1

• + Accurately identifies and transfers the male character's facial features and sunglasses.
• + Maintains the correct dynamic leg-crossing pose from Image 1.
• + Successfully incorporates the scarf from the character reference image.

• − Lower visual quality with noticeable artifacts around the feet and hands.
• − The hand on the left is anatomicaly distorted and missing fingers.
• − The transition between the scarf and the neck is messy.

Seedream 4.0

• + Higher overall image resolution and cleaner environment rendering.
• + Better preservation of the original hand pose and fingers from the reference pose.
• + Naturally integrates the clothing textures and colors.

• − Fails significant parts of the character reference, notably keeping the woman's long hair and face shape.
• − Mismatched character identity, looking like a hybrid of the two people rather than the character in Image 2.
• − The standing foot has an extra toe and poor anatomy.

GPT Image 1

• + Excellent anatomical details on the horse's musculature and hooves
• + Consistent lighting between the astronaut and the environment
• + Clean and cinematic composition with a realistic texture to the spacesuit

• − Failed to follow the specific spatial instruction for the horse to be 'on top' of the astronaut

Seedream 4.0

• + Vibrant and colorful space background with great nebula effects
• + Ddynamic lighting reflections on the astronaut's visor
• + High contrast and sharp focus on the horse's mane

• − Failed to follow the specific spatial instruction for the horse to be 'on top' of the astronaut
• − Perspective on the astronaut's legs and saddle area is slightly confused

GPT Image 1

• + Excellent preservation of the subject's unique facial features and skin patterns.
• + Very high visual quality with realistic lighting and texture integration.
• + Flawless adaptation of the coat and scarf to the subject's pose.

• − Failed to include the sunglasses and shoes mentioned in the prompt instructions.
• − The background is slightly simplified compared to the source image.

Seedream 4.0

• + Included almost all requested accessories including sunglasses, gold watch, and shoes.
• + Successfully captured the full-body outfit adaptation as requested.

• − Severely failed to preserve the subject's face, replacing it with a hybrid of the two men.
• − The overall image resolution and clarity are much lower than the source.
• − Poor blending around the edges of the person and the background.

GPT Image 1

• + Excellent fur texture and photographic lighting.
• + Clear 'TAXI' text on the driver's cap.
• + Composition feels intimate and fits the 'inside the taxi' prompt better.

• − The passenger's hand holding the phone looks slightly distorted.

Seedream 4.0

• + Successfully captures both the capybara and the businesswoman in a wider shot.
• + Good representation of a yellow taxi's exterior and interior simultaneously.
• + Captures the bored expression of the passenger well.

• − Text on the cap is garbled ('TAT' instead of 'TAXI').
• − The capybara's paws look more like human hands in gloves than animal paws.
• − The lighting is a bit flat compared to the cinematic quality of the first image.

GPT Image 1

• + Excellent typography with perfect spelling in all fields
• + Highly cohesive vintage gothic aesthetic
• + Cleaner composition with better readability

• − Missed the '7pm' time detail
• − Confused the 'Time' and 'Location' labels in the footer

Seedream 4.0

• + Successfully included all details including time, date, and location
• + Very detailed thorny border with literal parchment textures
• + Dynamic lighting on the jack-o-lantern and ground

• − Text on the small scroll banner is garbled and difficult to read
• − Composition feels slightly cluttered with the border encroaching on elements

GPT Image 1

• + Excellent preservation of the original facial features and expression.
• + The hair texture looks more natural and less uniform.
• + The transition between the hair and the forehead feels organic.

• − The shape of the head and hair mass is slightly asymmetrical.
• − A small portion of the glasses' frame appears slightly smudged near the hair transition.

Seedream 4.0

• + The hair is very thick and full as requested.
• + The lighting on the hair surface matches the ambient scene well.

• − The facial features were altered, making the man look younger and changing his eyes and nose shape.
• − The hair texture looks slightly like a wig or synthetic fiber.
• − The hairline is very sharp and blocky, lacking natural variation.

GPT Image 1

• + Excellent 3D miniature clay-like aesthetic with soft, refined textures.
• + Strong layout and composition on a clear diorama base.
• + Clean and readable text typography that integrates well with the scene.

• − The 'JAPAN' text is slightly off-center compared to the rest of the elements.
• − Minimalist approach might feel too simple to some users.

Seedream 4.0

• + Excellent PBR materials with realistic subsurface scattering on the fish and ikura.
• + Greater variety of sushi types provided within the scene.
• + Accurate text rendering and placement.

• − The text is placed right at the top edge, feeling a bit cramped.
• − The diorama base is less distinct as a 'raised base' compared to the other model.

GPT Image 1

• + Excellent caricature style with exaggerated facial features while remaining recognizable.
• + Successfully incorporates all elements: news desk, dog, and hockey items inside a cohesive illustration.
• + Captures the denim shirt from the source image accurately.

• − The watercolor/sketched style might be seen as a total redrawing rather than an image edit.

Seedream 4.0

• + Keeps the photographic style of the original person and background while applying caricature proportions.
• + Includes a variety of hockey references like the jersey and stick integrated into the scene.
• + Preserves the 'selfie' pose and environment of the original image better than the other model.

• − The hand holding the phone has severe anatomical issues (six fingers/clashing shapes).
• − The dog appears more like a standard stock photo insertion rather than a stylized caricature.

GPT Image 1

• + Excellent anatomical accuracy for all four animals.
• + Beautifully soft, realistic lighting and god rays that feel natural.
• + Clean composition with high-definition fur textures and clear facial features.

• − The 'tumbling' aspect of the prompt is less pronounced as they are mostly running.

Seedream 4.0

• + Captures the 'tumbling' and 'playful chasing' dynamic better with more active poses.
• + High contrast and vibrant colors create a very whimsical atmosphere.
• + Includes prominent dew sparkles as requested in the prompt.

• − Anatomical issues with the fox, which appears to have a broken or unnaturally twisted neck.
• − The dew sparkles look like floating white dots rather than natural light reflections on grass.
• − The tabby kitten's eyes and face lack the hyper-realistic detail seen in the other model.

GPT Image 1

• + Captures the specific character expressions extremely well, including the indignant look of the girlfriend.
• + Maintains the warm, golden-hour lighting associated with the Ghibli aesthetic.
• + Preserves the physical relationship and poses of the characters accurately.

• − Texture feels a bit more like colored pencil than a traditional painted cel style.
• − The background is quite blurry and lacks the 'scenic' detail typical of Ghibli films.

Seedream 4.0

• + Excellent watercolor texture that feels authentic to Ghibli concept art.
• + Strong character design that balances the source likeness with the anime art style.
• + Clean line work and soft pastel color palette.

• − The facial expressions are significantly softened, losing the humor of the original 'distracted boyfriend' meme.
• − Details of the man's shirt pattern are simplified compared to the source.

GPT Image 1

• + Excellent preservation of the subject's facial features and the dog's appearance.
• + Natural-looking hair motion that follows a consistent wind direction.
• + Added a dynamic fluffiness to the dog's tail and fur to match the wind effect.

Seedream 4.0

• + Larger, more colorful leaves add a stronger sense of seasonal atmosphere.
• + The hair blowing effect is well-integrated with the original hairstyle.
• + Good preservation of the overall background composition and lighting.

• − The leaf in the bottom left corner is very blurry and distracting.
• − The hair strands appear slightly more digital and less fine than Model A.
• − Added a lens flare in the top right that was not requested.

GPT Image 1

• + Excellent minimalist vector style with clean lines
• + Accurate spelling including the grave accent on 'Caffè'
• + Clear, professional typography

• − Ignored the 'light background' instruction, providing a black background instead
• − The cloche handle and steam integration is very basic

Seedream 4.0

• + Followed the 'light background' and 'subtle texture' instructions perfectly
• + More artistic and detailed cloche design with better steam representation
• + Stronger 'vintage' feel with the cream and brown palette

• − The accent on 'Caffè' is slightly misplaced or looks more like an acute accent
• − The banner folding logic is a bit inconsistent at the edges

GPT Image 1

• + Excellent adherence to the muted navy and cream NASA-inspired palette.
• + Clean, professional flat-vector icons with very crisp lines.
• + Accurate and legible typography for the names and main labels.

• − The layout for the steps is unorganized and doesn't follow a logical flow.
• − Contains significant typos like 'EARLLUNAR' and misalignment between text and icons.

Seedream 4.0

• + Logically follows the number of steps (1-6) in a clear visual sequence.
• + High-quality rendering of the Lunar Module on the surface.
• + Includes a trajectory arc as requested in 'Translunar' step.

• − Text rendering is messy with typos like 'Surfcce' and overlapping characters in 'Landing 6'.
• − The flat-vector style is inconsistent, mixing flat icons with more detailed 3D-shaded objects.
• − Layout is a bit cluttered at the bottom.